Marius
Meiners

For twenty-five years the answer to "how do I get found" was Google, and an entire industry - search engine optimization - grew up around gaming its rankings. Then, in late October 2024, OpenAI bolted a search engine onto ChatGPT, and Meiners had the specific, slightly greedy thought that founders have and most people don't: the whole apparatus was about to reset, and there was money in being early to the reset. "I realized, hey, this is actually quite a massive change in how consumers find things, find products, find services," he told Inc. The established business models, he figured, would re-emerge in an AI-first world. Somebody would need to sell the SEO of the ChatGPT era. The industry even had a name ready for it - generative engine optimization, or GEO.

Peec AI is the dashboard for that discipline. You feed it the questions your customers might ask an AI. It runs them across ChatGPT, Perplexity, Gemini and the rest, then reports back: which brands got mentioned, how often, in what order, and whether the model spoke about them warmly or dismissively, scored on a scale. It is, in the plainest terms, a mirror. Brands look into it to see the version of themselves that a language model describes to a stranger, and then they pay to make that reflection more flattering.

The customers who pay are not obscure. Peec's roster includes Chanel, the publisher Axel Springer, the voice-AI company ElevenLabs, the automation tool n8n, and the travel giant TUI. As of late 2025 the company counted more than 1,300 brands and agencies and was adding roughly 300 a month. What is unusual is not that these companies want the data. It is that two years ago the thing the data measures did not really exist.

There is a tidy recursion at the center of all this. Peec AI helps companies get recommended by AI search engines, and roughly one in five of Peec's own new customers arrives having found Peec through an AI search engine. The product is, in a modest way, its own best salesperson. That is the sort of detail Meiners seems to enjoy, because it is both true and slightly absurd, which is a fair description of the whole category.

When Meiners walked into Antler's Berlin accelerator in September 2024, he had the profile of someone who might never ship anything: no co-founder, no fixed idea, and, by his own count, four years since he had last written a line of code. His first pitch, three weeks in, collapsed when the relevant legislation changed. He and an early partner, Daniel, then cycled through legal tech and a regtech idea, killing each within two or three weeks using a single rule - market pull over supply creation. "People really don't want to buy stuff they don't really urgently need in a B2B setting," he has said. The way he tested for urgency was not to pitch but to ask open questions, so he could tell whether AI search was genuinely the thing keeping marketers up at night. When the best SEO people he knew kept raising it unprompted, he had his answer.

The next decision is the one that gets retold. Rather than wait for a technical co-founder to build something, Meiners sat down with V0, Vercel's AI coding tool, and, as he puts it, "just vibe-coded it in a day and a half." The prototype accepted a batch of user prompts, queried the language-model APIs, pulled out the brand mentions, scored the sentiment from one to a hundred, and displayed the result. That core loop is still what Peec does today. He built it despite not having coded in four years, which is either a statement about Meiners or about the tools - probably both.

Then came the part that separates a demo from a company. Before writing any production code, he took the V0 prototype and used it to sign eight letters of intent. He sourced the prospects from people already arguing about AI search on LinkedIn and X, which made them warm before he ever said hello. His pitch was a flattery of the buyer's own instincts: wouldn't it be nice to tell your board, or your CMO, how you are performing in this brand-new channel? He is clear-eyed that a letter of intent is a soft thing - "legally non-binding," he notes - but it beat a nod in a meeting. What it did not beat, in his ranking of proof, was money. That distinction runs through everything he says.

Antler put in about 100,000 euros in December 2024. Meiners partnered with Tobias Siwonia, a serial founder who became CTO, and Daniel Drabo, whom he calls "a hustler," as chief revenue officer. Siwonia built the real product in roughly six weeks. Peec AI launched in February 2025. For the first three months the founders took no salary, worked out of Antler's office, and ate two-euro frozen meals while they waited to find out whether their bet was right.

The logic was that a hundred-to-two-hundred-person company needs to know how it looks to ChatGPT just as urgently as a global brand does, but cannot stomach an enterprise procurement cycle to find out. So Peec priced its entry tier at 85 euros a month, later building tiers up to 425, and swept up a mid-market that better-funded rivals had ignored - eventually more than 2,000 of them. His read was that AI search would follow the same shape SEO did: "SEO as a topic is so relevant for so many small customers," he has said, betting on breadth rather than a handful of enterprise logos.

It is a very particular kind of contrarianism - not the glamorous sort, but the sort where you accept a less impressive customer with a more urgent problem. That fits a broader thing Meiners repeats: "Try to be effective rather than trying to be cool when it comes to business fundamentals." He applies the same test to the advice he disagrees with. Being scrappy at the start, he says, is essential; the mistake is staying scrappy. "You got to shift into deploying a lot of capital very quickly when you find product market fit." Knowing which phase you are in is, in his telling, most of the job. He went from two-euro frozen meals to raising eight figures within a year, which is a fairly literal demonstration of the point.

The detail that people fixate on is that Meiners was, as a teenager, one of the hundred best League of Legends players in the world. He played tournaments, streamed on Twitch, and lived inside the game's ranked ladder in the way that only a certain kind of teenager can. He credits that period, rather than any classroom, with teaching him what excellence actually requires - the unglamorous, repetitive climb toward mastery that transfers, he thinks, to any domain including running a company.

The classroom, by his own account, went less well. He describes his original credentials bluntly as "very, very shit," redid his schooling, and studied economics at what he calls Germany's lowest-ranked university because he could not get into better ones. From there he assembled an unusually complete view of startups - as an employee, an investor at a small VC, an M&A adviser at PwC, and an LP - before deciding to sit on the other side of the table and build one himself. It is the sort of career that looks like drift until it suddenly looks like preparation.

He is, he says, a private person - "I don't even have TikTok or Instagram" - which sits awkwardly against his own advice that "if you want to win in B2B SaaS in 2026, you have to be on social, you have to love content." He has resolved the tension the way an effectiveness-first person would: he forces himself onto LinkedIn because the distribution demands it, whether or not he enjoys it. Away from the company he plays the occasional video game, belongs to a card-gaming community in Berlin, and turns up to the city's Sunday day-club scene. One of the books he recommends is Eckhart Tolle's "The Power of Now," which he has said helped him through the harder stretches of scaling.